INDEX

Explanations

proper nouns that refer to specific entities or individuals

oai_token-act-pair · gpt-3.5-turbo

instances of the word "last" indicating finality or the end of something

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ingen

-0.78

amber

-0.76

kaya

-0.69

velt

-0.66

gans

-0.65

houses

-0.65

aband

-0.65

ancel

-0.65

adders

-0.64

obs

-0.64

POSITIVE LOGITS

 gasp

1.05

decade

0.91

 ditch

0.88

ebin

0.85

 month

0.84

 supper

0.83

 rites

0.82

 week

0.82

 semester

0.80

 installment

0.80

Activations Density 0.051%

proper nouns that refer to specific entities or individuals

instances of the word "last" indicating finality or the end of something

No Comments

No Known Activations